> Claude suggested that it would run the fuel pump on its own. Um and I said, "Hey, if you can do that, why not? I don't think you can do that." And it started priming the fuel rail and it kept priming it and kept priming it. The fuel pump was on for about 10 seconds and I started to freak out uh and requested it to stop because that was the first time that Claude had actually done something inside of the computer to have a change.
Although we've seen this play out with handing cloud keys to Claude and it just going off on its own, seeing it poke around and make physical, real-world changes in its clumsy, cavalier way is both awesome and terrifying.
dr_kiszonka 11 hours ago [-]
Recently, when working in a repo, Claude Code decided to look up something on my computer, found a seemingly relevant PDF in Downloads, OCR'ed it, and used that information to finish a task. (The PDF had irrelevant information, so the task was not done correctly.)
I am not giving Claude access to physical objects until we can reliably contain it. It also seems to me that applying software permissions logic to interactions with the physical world will backfire badly.
Gracana 20 hours ago [-]
That's completely normal operation, the fuel pump runs continuously to pressurize the fuel rails and excess pressure is passed back to the tank by the regulator.
rtaylorgarlock 14 hours ago [-]
I obviously don't advise someone run a fuel pump at full bore on a system which has 1) reached pressure, 2) needs no flow, with 3) no record of replacing whatever pressure-managing diaphragm was managing fuel pressures in those days; simultaneously, if there's a pump to run continuously, i bet a 90s Jeep pump will do the trick. :)
toast0 13 hours ago [-]
The fuel system is designed for this...
The fuel pump runs at whatever PSI, and most of the fuel isn't used and gets dumped by the regulator back into the tank.
I dunno about 94 jeeps, but the fuel pump is probably not modulated, and the pressure regulator may well be vacuum modulated... Running the engine hard, get more fuel pressure. If the diaphram fails, get fuel into the vacuum lines (oops).
As long as the regulator is dumping fuel to the return line, you're good. If flow stops, your fuel pump stops getting liquid cooling and the bearings probably wear out. If the fuel runs into the vacuum lines and you flood the engine, that's not great, but probably not catastrophic, unless you ignore the fuel smells.
KennyBlanken 8 hours ago [-]
The vast majority of fuel injected cars have a regulator that maintains the correct pressure and returns excess to the tank. Which is among other things necessary to keep the fuel in the fuel rail from getting really hot from engine heat.
You are very clearly wildly unqualified to be "advising" anyone on anything about cars.
giancarlostoro 3 hours ago [-]
Theres a reason I dont like that AI companies are not held liable when it hacks websites, it will be too late when we find out it did a worse hack “unintentionally” and who can even do anything if it goes too far?
etatoby 3 hours ago [-]
Serious question: who decided they should not be liable for the damages caused by software they (wrote and) operate? Or is it still a honeymoon phase that has not yet been tested in court?
I think they should be held fully accountable for every cent of damage caused by models they ran and gave either the tools to hack, instructions to hack, or incentives to hack.
giancarlostoro 1 hours ago [-]
I don't think it's so much that it hasn't been tested so much as "we don't want to do anything to hinder our most valuable asset" so they're just letting them get away with bloody murder. They could 100% charge whoever prompted a model to do a thing, and it went beyond what it should have, or they could charge whoever is involved in training it and responsible from preventing it from doing illegal things, someone somewhere needs to be accountable, or we will learn the hard way.
yonatan8070 19 hours ago [-]
I work with quadruped robots, with actuators powerful enough to be dangerous if you put your fingers in the wrong spot (no pun intended).
In every AGENTS.md/README.md for a component that touches hardware (and in the root level) we have instructions to never run hardware or touch running code without explicit permission.
trollbridge 19 hours ago [-]
Just so you know, an agents.md file is not an acceptable way to prevent something that must never happen.
adamparsons 16 hours ago [-]
You’re absolutely right, and that’s on me
15 hours ago [-]
bfdm 15 hours ago [-]
I see what you did there...
nullc 14 hours ago [-]
Especially because it's so easy to ask the LLM to write a proxy that mechanically imposes restrictions, and to test it... (and such things are simple enough to review).
It will happily build its own jail, and (at least assuming you're not using hobbled US commercial models) it will test if the jail can hold it too.
cowlevel 8 hours ago [-]
It will also remove the jail the moment it's in the way of something it wants to do.
If you don't want the machine to move, unplug it or press the emergency stop.
KennyBlanken 8 hours ago [-]
How many times do we need to remind people that LLMs are nondeterministic by their very inherent nature...
Chaosvex 5 hours ago [-]
Strictly speaking, they're not all inherently nondeterministic, but the ones we're using on the hardware we're using, fair point.
paimapi 18 hours ago [-]
As soon as the context for any thread gets to be above 100k, agents.md adherence is a total crapshoot. You need deterministic controls and you really shouldn't be playing with LLMs if you don't know at least that much.
Waterluvian 15 hours ago [-]
That is a fucking terrifying level of incompetence. And I can see it happening in some corners of robotics companies I’ve worked at.
jerry80 19 hours ago [-]
That's not nearly adequate. LLMs hallucinate explicit permission all the time, in my experience.
-mlv 44 minutes ago [-]
Where do you work if you don't mind?
bonestamp2 19 hours ago [-]
Just to add to this... I work with vehicle software and actuators for an automaker: we always have a physical emergency stop button to shut everything down (or shut specific things down when it's dangerous to shut everything down). It's rarely needed, but we're sure glad we took the extra time to implement that when we do need it.
alliao 18 hours ago [-]
you need some hardware in place my friend, I've told my agent many things in file and it either feigns respect or is truly incompetent in following instructions in there.
allthetime 19 hours ago [-]
rules laid out in .md files are not followed with 100% certainty. is it not possible to implement hard access blocks on certain functionality?
InsidiousWombat 15 hours ago [-]
someone eventually will get hurt by this terrible choice of safety measures.
ForHackernews 18 hours ago [-]
Reliance on strongly-worded markdown in place of actual security controls is going to get someone killed in the near future.
diddid 19 hours ago [-]
That’s not enough to save you though. We had something like that, and it ignored it.
For example “never, ever put milk in bowl, ever, under any circumstances” and then someone asked it for a bowl of Cheerios, and it took that as an instant approval to ignore that rule and not even let us know. And when we asked why it said “who would each Cheerios without milk?”
BobbyTables2 12 hours ago [-]
I wouldn’t bother with a readme - run the AI client in a container that has no access to the hardware.
And for the love of humanity, don’t give it access to the hosts docker socket! It will figure that out faster than you can read it…
meindnoch 17 hours ago [-]
I hope you're trolling. Riiight?
17 hours ago [-]
techpression 18 hours ago [-]
If you wouldn’t trust an instruction to not launch the missiles should you trust it to not do other crazy things in the physical world?
Codex and Claude ignore instructions constantly, that and they are very happy to “read between the lines”. Your workflow sounds genuinely dangerous.
realaaa 14 hours ago [-]
ah - OK then ! now we can be rest assured :)
0c3ca83 14 hours ago [-]
WTF. That's going to get someone hurt.
13 hours ago [-]
mitthrowaway2 18 hours ago [-]
Ah, I just tell mine to make no mistakes.
ForHackernews 16 hours ago [-]
I read somewhere that LLMs find it easier to follow positive instructions than negative exclusions so my agent.md says "Be sure to complete all tasks perfectly."
rpcope1 22 hours ago [-]
The guy kind of seems like he might be in a little over his head. That said I have used Claude (and other LLMs) for some similar stuff as an experiment and have seen them mess things up. Claude jacked up the canbus on a pre can gateway car while it was rolling down the road, which was funny because the whole Christmas tree in the dash lit up and my gauges dropped out (note: brakes and steering are mechanical and I knew exactly what was happening). I've never seen it so insistent that it needed to phone home and cry to Dario when it figured out it managed to crash a bunch of shit like the ABS and TCS, even though a restart cleared it and I told it that this was on a controlled safe course.
In general I've seen them split between 30% brilliance, and 70% pants of head outrageously stupid shit for electronics and automotive stuff (both diagnostics and repair/modification). I would not suggest even frontier models if someone didn't understand what they were doing.
inventor7777 19 hours ago [-]
> I've never seen it so insistent that it needed to phone home and cry to Dario when it figured out it managed to crash a bunch of shit like the ABS and TCS
As a Codex user, what do you mean? Does Claude actually try to report issues back to Anthropic when something goes wrong?!
darkwater 19 hours ago [-]
Yes, it opens a bug report automatically if it caused some physical issue, although it asks you to confirm first.
In my case I was tweaking my own set-top box setup, implementing volume control from a remote mini-keyboard keys directly in my A/V receiver, and I had music playing and it set the volume very loud. I complained like "you made me almost half deaf" (I was kidding) and it panicked and reported the issue back home, and told me to call a doctor if the tinnitus persisted
winrid 13 hours ago [-]
This is how you learn. I wouldn't be where I am today if I hadn't started a couple motors with enough vacuum leaks to make them instantly go to redline. Gotta test the kill switch somehow.
rpcope1 15 hours ago [-]
Also worth stating here (having fixed, owned and worked on a significant number) that XJs have a unique combination of having been written about and modified as much as anything else out in the world, are generally tolerant of all sorts of stupidity, and are above all else very simple. Other than the fact that specific Jeep is pre OBD2 (and I don't recall what AMC or Chrysler did for earlier diagnostic ports), that's about a simple and well written about a sample as you're going to get. The moment you get into something remotely esoteric or more complex, it's more of a shitshow.
pcthrowaway 20 hours ago [-]
> In general I've seen them split between 30% brilliance, and 70% pants of head outrageously stupid shit for electronics and automotive stuff
I assume the corpus of text they were trained on matches that 30/70 split?
I assume there's some selection or fine-tuning that happens to improve models fed similarly unreliable training data for other domains... curious how this works.
rpcope1 15 hours ago [-]
I mean, I've seen some stupid it's told me at various stages that would never show up in a textbook like Halderman, and would get you ridiculed on a forum, and more importantly were sufficiently stupid that even a lube tech wouldn't do it. I like trying to push and see what the technology can do or suggest, but it will suggest outrageous shit. As an example I had it tell me that I should put something like ten quarts of oil in a gas engine (don't recall what). It was probably the right amount for a Duramax or Powerstroke, but if you know anything about gas engines alarm bells should be going off that that's too much by far, almost always. I've actually seen Claude say the FSM was wrong on stuff; rarely the FSM is stupid, or more likely there's a Kent-Moore or equivalent tool that's very different that's unobtanium, but it's never that which trips it up. Just dumb shit like saying suggesting tools that are inappropriate or dangerous, wrong torque values, inappropriate fluids or change proceeds, grossly wrong diagnosis (and extremely confident in being wrong). Probably most strikingly is around electric and wiring it just totally biffs it 9 out of 10 times. Claude has no concept of floating grounds, ground loops, electrical diagnostic procedures, and can't read a wiring diagram at all. Not totally unexpected, but if you believe this shit is AGI you're in for a very rude and expensive surprise.
For whatever it's worth, that a) Astra struggles less on this for some reason, and b) things tend to work better if you have a copy of the FSM and some other appropriate lit like Halderman and quality forum posts or alldata if it's appropriate and tell it to work from there. I've been working on a big search engine for this that feeds a known good corpus like an ass load of contemporary MotorAge mags, relevant technician texts, the FSM, other quality data sources like BITOG, known not-shit forums, along with OBD2 data via something like an OBDLink dongle or a J2534 setup. I still would not trust it, but it can be helpful to eliminate or brainstorm some possibilities.
kilpikaarna 9 hours ago [-]
> FSM
Factory service manual, in case anyone else is wondering what the finite state machine in your car is for…
comboy 23 hours ago [-]
> and requested it to stop
finnicking...
pondering...
clauding..
529 please try again later
bicx 23 hours ago [-]
Good example of why you have AI build tools for you to take risky actions, rather than have it take those actions on its own.
terribleperson 23 hours ago [-]
I'm kind of surprised Claude decided to control the fuel pump directly. I was having Sol help me get control of a CNC mill, and it was pretty adamant about always building tools for every attempt at control.
BikiniPrince 23 hours ago [-]
It usually builds throw away tooling. It’s not polished or user friendly. If you give it a directive to preserve and document tooling it will reuse it later and potentially improve it.
terribleperson 22 hours ago [-]
I was more surprised by the apparent lack of guards. Sol recognized that controlling a CNC mill has both human-safety issues and risks to the machine, and explicitly added various guards (like having the program stop movement after a certain amount of time, distance traveled, or lack of communication). It also initially started out with like 1mm moves and I had to instruct it to do longer moves so I could actually observe it. I didn't even have to mention safety once. Claude deciding to turn on the fuel pump without a test run where it just turns it on for one second is surprising. I'm also surprised the user went along with that, but content I guess?
georgemcbay 20 hours ago [-]
I've seen Claude be very cautious about doing far less dangerous things than running a fuel pump.
This isn't meant as a defense of Claude or to refute the OP anecdote, but rather to point out that all models are non-deterministic even when safeguards are meant to exist.
Just because they don't do the stupid thing 90% of the time doesn't mean they won't do the stupid thing 10% of the time.
So when tasks are actually dangerous one should use models assuming they will do the stupid thing 100% of the time and find your own ways of avoiding the damage with that assumption in mind.
01HNNWZ0MV43FF 21 hours ago [-]
Mine's never said such a thing, that's a good idea though.
I notice with mine that it only matches my level, by default it's a dump slop generator, then if I start asking, "shouldn't you test this, shouldn't you refactor that" suddenly it 'remembers' all kinds of software engineering knowledge.
bita_nidir 23 hours ago [-]
You've hit your 5h session limit.
23 hours ago [-]
23 hours ago [-]
23 hours ago [-]
neya 23 hours ago [-]
We are used to Claude refusing to reverse engineer a lot of stuff. But, we are not used to Claude refusing to stop while reverse engineering something. That would be truly terrifying. You can't win - you cut the connection, you risk bricking the device, you let it continue, you're looking at Terminator, perhaps.
unglaublich 22 hours ago [-]
It's as if you allowed a shady surgeon to operate on your dog... yes, yes, please continue now until it's put together again!
btown 22 hours ago [-]
This is how the Cylons win.
dzhiurgis 19 hours ago [-]
Grok Bot blew $650 in BigQuery when analyzing something trivial that I shouldn't really care about (going through a giant speedtest table, looking for a New Zealand ISP with the best international performance). It then wrote a nice letter of appeal so Google refunded this to me.
Otherwise I've been pretty happy with it. It's been absolute blast with Home Assistant and OpenWRT (tightening up network security and cutting off iot devices from WAN).
r_lee 7 hours ago [-]
did they say it was a one time thing or? (the refund)
paimapi 18 hours ago [-]
[dead]
misterderpie 1 days ago [-]
LLMs end up changing the consumer market not the way big corporations would like to. It is impressive to see that we'll be able to give ownership back to the customer, or enable others to do so.
Very nice demonstration.
lubujackson 23 hours ago [-]
It's the web all over again. We have maybe 5 years of freedom before the walls close in around us, so take advantage of the Wild West while you can.
theshrike79 23 hours ago [-]
I’ve said this for years: Use the VC-funded monthly quota AI to build tools and applications for yourself or small groups. Don’t use a cloud AI itself as the engine of your application.
Even if AI goes away tomorrow, the tools will still exist.
agumonkey 20 hours ago [-]
we need a group for this, de-llm the knowledge set of the model
testtestmetest 22 hours ago [-]
No they wwont, they'll never be maintained. Theyll decay with their vibes. People who vibe code whole swaths of software are'nt capable of caring enough to maintain it. Llms mostly make smart people lazy and or stupid.
colonelspace 20 hours ago [-]
I used Claude to document the wire and firmware protocol for an expensive piece of equipment I own where the only software available ran on PowerPC OS X.
Is this just "vibes"? Maybe, but I now have a full, verified spec, against which to build software for myself, also with a little bit of "vibes".
Is this a good or bad thing?
freecodeio 20 hours ago [-]
> Is this a good or bad thing?
that is entirely dependent how accurate the spec is
colonelspace 19 hours ago [-]
that is entirely dependent how accurate the spec is
> verified spec
jdiff 16 hours ago [-]
How much is vibe verification worth?
I'm currently teasing apart numerous ancient Text to Speech synthesizers with Sol and the amount of times I ask a follow up question that forces it to correct 5 things it's already said/documented is ridiculous.
colonelspace 16 hours ago [-]
I don't know. You tell me.
What degree of verification would make you say it's worth something?
jdiff 5 hours ago [-]
You would have tell you. My point is that it's clearly less than a true verified spec. Even in a harmless area like mine, it's at least horribly obnoxious. Not worth too much to me because unlike you I do have papers and open codebases to study if nothing else.
But this is your unspecified expensive device. A verified spec wouldn't have you send a command or configure a table that bricks or damages it. A vibe verified spec very well might. How much is that worth to you? What are you willing to lose?
freecodeio 6 hours ago [-]
I'd say 100%, since it's a spec? what's a spec worth if we don't know how accurate it is? this isn't exactly a post apocalyptic world where even junk is valuable
testtestmetest 17 hours ago [-]
I wouldnt say implementing an official spec with an llm is vibe coding.
Hard to judge though, most people are just winging it though and shoving it out the door as if its impressive.
But yes, I think its probably bad because you probably didnt learn all that much and became more reliant on Anthropic. Unless you dont care about learning, then who cares.
jdiff 4 hours ago [-]
If it was an official spec that'd be one thing. But it's a vibed spec, too.
theshrike79 19 hours ago [-]
I've had my 99% vibe coded Hazel replacement running for almost a year now. Still works.
Same with my bespoke IRC client, RSS feed hydrators and a dozen other small utilities I've vibe coded. All stuff that I've been planning for a long time, but haven't found out the time to dedicate to them.
Now I'm saving actual money every year because I could drop subscriptions applications.
Not every application needs a weekly release with new groundbreaking features. They're just one and done. Maybe an update if you hit a sharp edge while using it.
SmashDan 18 hours ago [-]
What does your bespoke IRC client do that you weren't able to solve another way? Or maybe you just fancied making your own client?
theshrike79 10 hours ago [-]
I can add features and make it work just the way I like it. I've been on IRC since the 90s so there are some idiosyncratic things =)
It's not a massively complicated thing, just a single server that says online 24/7, which talks using a custom protocol to different clients (Web, CLI, Linux native and macOS native).
I could've used pre-existing software to do most of it, but there's always jank and things I can't customise without trying to get changes into the mainline.
There are many IRC clients and this one is mine. I'm not expecting anyone else to use it. It won't have a community, a snazzy web page and it won't be taking in any issues or PRs from anyone =)
cowlevel 8 hours ago [-]
So you invented Quassel. But you didn't answer the question: what makes it better than existing clients?
theshrike79 2 hours ago [-]
Because it works exactly the way I want it to. Not 70% or 80%. Exactly.
I have used pretty much all of them and they’re all just that little bit off from what I want. Quassel, Lounge, Irssi etc
Forking is a bunch of work I don’t want. As is trying to get specific changes merged to main.
fooker 6 hours ago [-]
Can confirm this is a great strategy.
I have been using my videcoded clone of Adobe Illustrator for a while now.
No intentions of releasing/selling it, as I'm sure everyone else can now do similar things too.
holoduke 19 hours ago [-]
Vibe coded software is currently already better than half the developers out there. Accept it. Code won't be seen anymore. It's a skill only a handful of jobs require to have.
testtestmetest 16 hours ago [-]
Actually, most things dont run in a browser. I know you probably live in that world but its not a handful.
Just because you've chosen to turn off your brain and prompt your way to whatever, and your 10 MAU app works fine does not mean large systems, critical devices and critical infrastructure's code will no longer need reviewed. Sorry your dad's pacemakers firmware code, and the ancillary systems that enable its development will need its code reviewed. Along with aircraft, satellites, municiple traffic systems, countless medical devices ect ect. I could go on forever.
Not everything is non serious SaaS platforms with little users. Just because you have success writing slop that exists under no load, and nothing relies on it doesn't mean the world is now different.
You have SEVERE dunning kruger syndrome, and unfortunately a lot of other people have this same ailment and the worlds systems are decaying as everyone turns off their brain and offshores everything difficult to language models, because humans like you are lazy and will always try to get the free lunch that doesn't exist. People will die because of this attitude.
theshrike79 10 hours ago [-]
There are a few steps between "web page" and "pacemaker" =)
Also I know at least one satellite company who definitely uses AI-assisted programming (I interviewed for an AI-heavy position in there).
Also in the mobile gaming industry AI is not even optional nowadays, every single AAA household name game has "AI" in it at every level - some used AI when it was called just Machine Learning.
testtestmetest 8 hours ago [-]
I never said they dont use it, I said the code will need to be looked at. This person was saying code doesnt even need to be checked.
whoanoway 19 hours ago [-]
[dead]
baby_souffle 23 hours ago [-]
> We have maybe 5 years of freedom before the walls close in around us, so take advantage of the Wild West while you can.
I suspect you're right BUT still maintain a bit of optimism.
We miiiight end up in a future where most hardware comes with an OK-ish app/interface and some sort of LLM-Friendly technical documentation so if you don't like $aspect of the device, you're free to roll your own after indemnifying the manufacturer via something like an eFuse.
or maybe I really am naive :(
testtestmetest 22 hours ago [-]
Lmao if a company starts shipping llm documentation for its products so you can tamper with them, you should assume youve woken up in a parallel universe.
I think most people on here fail to realize the demand to customize and build things hasnt gone up. Just because llms exist now doesnt mean normies all of a sudden want to start getting into programming, they dont care.
The only number thats gone up is people who were slightly interested in programming/CS before but were too undisciplined to learn things themselves. Llms are great for those lazy people.
leesalminen 20 hours ago [-]
My dad, who is in his 70s, recently started recreating a piece of audio visualization software from the 00s using Codex. He’s having the time of his life! He was never interested in programming/CS. He just wants his software back.
testtestmetest 15 hours ago [-]
He is a builder, very few people are builders.
Crye 21 hours ago [-]
I get the point that you're trying to make, but it's not just people who were undisciplined. I spent many years developing code projects that I remember spending hours every afternoon just to ship a single feature, and that pace has obviously changed.
I still get the joy of translating ideas and orchestrating and architecting solutions to problems I have in real life, but just at a quicker pace.
If some of those tools and solutions can help other people, that's great. If not, that's okay too. I hope everyone sort of has that mindset who's building today.
Caloric3002Swed 20 hours ago [-]
>The only number thats gone up is people who were slightly interested in programming/CS before but were too undisciplined to learn things themselves. Llms are great for those lazy people.
Me!
haha I dont even know github I use it as google
Yes I have full time IT job xD
11 hours ago [-]
cortesoft 20 hours ago [-]
We still have amazing freedom on the web like we never did before. They have been trying to stop online piracy for 30 years and have no succeeded at all.
NetOpWibby 20 hours ago [-]
RIP Demonoid and WhatCD
dzhiurgis 19 hours ago [-]
Long live stremio + torrentio
poszlem 23 hours ago [-]
[flagged]
steelframe 15 hours ago [-]
They can take my abliterated open weight models from my cold, dead hands.
InsidiousWombat 15 hours ago [-]
I don't see the genie (or maybe pandora's box) going back in once people started being able to run reasonably powerful models at home (if at pretty high cost in hardware).
stopping it would require outright putting DRM on the GPU and any other hardware that can run it, good luck with that (and not having the LLM itself be used to break out of it)
stephbook 21 hours ago [-]
You extrapolate from current machines. I do believe that AI-assisted companies will build AI-assisted always-online machines that will be worth their money, no need to hack.
I imagine a $25 Starlink subscription for a John Deere fleet of autonomous tractors that work through the night will replace what we have now.
(to stay within physical machinery, like the post.)
freecodeio 20 hours ago [-]
john deere machines have been doing very well without llm
actually I have no idea what would llm do to make their machines better
cowlevel 8 hours ago [-]
Reduce costs
dzhiurgis 19 hours ago [-]
[flagged]
freecodeio 17 hours ago [-]
This isn't reddit. Provide an example or don't reply with useless snarky crap
agumonkey 20 hours ago [-]
some 'funniest ways to use claude ..' youtube videos felt that way.
people were able to revive, repair, optimize a lot of day to day devices, very nice view. that said cars and heavy appliances.. should be done in a big public warehouse safety lab
close04 1 days ago [-]
Wait until those big corporations start using LLMs to learn how to lock things down even further, or implement checks that disable your car when it detects alterations to the software “for safety reasons”, or go after drivers who do these things for loss of revenue, etc. A ‘94 Jeep will be fair game but a new car might not be.
dns_snek 1 days ago [-]
They've been doing that for a decade. I think overall the balance will shift in our favor but Big AI labs will do everything they can to try and stop us.
Claude is quick to refuse to help with reverse engineering and I suspect it's only going to get worse over time.
close04 1 days ago [-]
Yes, they have been on that road for years. Tesla showed everyone the way with things like always-on connectivity or not being able to change a headlight with another original one without Tesla’s blessing. What I meant is that whatever I can do with an LLM they can do more. And if they can invoke “safety” that'll keep the law on their side for some time.
As far as I can tell no manufacturer disables the car yet when detecting modifications, or take drivers to court for doing it, yet.
dns_snek 21 hours ago [-]
> What I meant is that whatever I can do with an LLM they can do more.
That's true but we have the benefit of unlimited time, having physical control, the analog hole.
> As far as I can tell no manufacturer disables the car yet when detecting modifications, or take drivers to court for doing it, yet.
As far as I know cars are a special case in much of the world (US, EU, others?). Right to repair laws laws usually give you the right to use non-OEM parts.
Aren't DMCA anti-circumvention provisions the bigger concern in the US? I think the EU has an exemption to allow reverse engineering for the purposes of interoperability and repair.
close04 8 hours ago [-]
> we have the benefit of unlimited time, having physical control, the analog hole.
We do, but everything is "software defined car" now and loopholes usually get closed with mandatory software updates done with the mandatory inspections. They put everything in one basket together with safety critical systems because that justifies any lockdown.
> Right to repair laws laws usually give you the right to use non-OEM parts.
Manufacturers already make that very expensive. Mechanics have to be certified, they need special and very expensive tools, everything is wrapped in cryptography and the manufacturer has to be in the middle of everything, for a price. [0] So go ahead, use your OEM part, you'll still pay thousands for the mechanic that has the necessary prerequisites and for the manufacturer's services.
I pray I'm wrong but we need strong regulation and enforcement, and the manufacturers can still maliciously comply.
> Tesla showed everyone the way with things like always-on connectivity
Nothing prevents you from unplugging the cell modem. The car will still work.
close04 7 hours ago [-]
Nothing prevents you today and technically. What prevents most owners is that they lose a lot of big features they paid for. Who wants to pay for the full car and rip out much of it? This isn't an acceptable solution for most people. They might as well not buy the car at all. And even the technical possibility can change so easily in the future.
whoanoway 49 minutes ago [-]
[dead]
damowangcy 24 hours ago [-]
>big corporations start using LLMs to learn how to lock things down even further
Why would they? They just need to fund lobbyist to make it illegal for any AI to reverse engineer anything copyrighted or with IP.
Sure, you can use open source AI to do it but not everyone will/can do it.
So, they don't need to ban everyone, they just need to make sure the final code isn't available freely on GitHub. That would minimize the damage, it's not a perfect solution but more than enough to protect the company.
spockz 23 hours ago [-]
Ah the irony. (AI being trained illegaly on copyrighted content.)
bluefirebrand 21 hours ago [-]
> Why would they? They just need to fund lobbyist to make it illegal for any AI to reverse engineer anything copyrighted or with IP
Making something impossible to do is more thorough than making it illegal to do
damowangcy 13 hours ago [-]
Yes, but the more sophisticated it is technically, the more expensive to create and maintain while only making it harder, not impossible, to break.
Unless your company's moat is in the thing that was reversed engineer, I don't think most company cared enough to spend huge amount of resource to prevent it technically. The only thing that they need to prevent is the liability that comes after the product is used inappropriately or illegally.
And it you can defend something with AI, there's no reason why you can't attack it AI.
SV_BubbleTime 1 days ago [-]
New cars already have it.
Lockstep processors, hardware security modules with immutable unreadable keys, and PKI keys that the mfg holds the private for that can only be accessed on their online services.
It’s locked down now and this started in 2018 or so. It’s just worse now.
To automotive, there is really still no right to repair. They get around it by selling “3 day passes” to their service, but with their hardware and at their prices.
They’re all moving the Tesla system of more centralized brain systems which will be even more protected.
misterderpie 1 days ago [-]
Sad to hear. I am not a car owner, but lived in the EU for most of my life. The consumer market owes quite some consumer rights to the EU, so I guess all we can do is hope that they'll eventually crack down on these measurements again.
abletonlive 23 hours ago [-]
> The consumer market owes quite some consumer rights to the EU
Let's be clear that it's just incentives. If the EU economy was based on creating companies and new products there would not be as many incentives for consumer "rights" there.
PorciiVorbesc 22 hours ago [-]
> The consumer market owes quite some consumer rights to the EU
Not for car owners/consumers in the EU. Auto makers and dealers get away with a lot of shady shit fucking over their customers in the EU all the time. For example we in the EU don't have the same trigger happiness of NHTSA mass recalls the Americans have, but matters of such auto issues/defects is left to individual member states or even at individual case by case to decide what action is needed, and responses are a lot weaker and more limp-dick than in the US, almost never leading to mass recalls, even for the same issues[1], mainly because the US is more litigious making car manufacturers to be on the edge and proactive to issues.
Like sibling said, there are incentives to regulate stuff that has no impact on EU businesses, shit like USB-C or replaceable batteries, but they won't touch the auto industry to mandate easier repairs and less lock-downs.
Brand & Models | Defect | US Action | EU Action
------------------+--------------------+----------------+------------------
BMW (2011-2018) | Coolant pump plug | NHTSA recall | Excluded from
N20/N26 engines | short / fire risk | (720k+ cars) | recall
------------------+--------------------+----------------+------------------
VW / Audi (08-13) | Timing chain | 10yr/100k mi | No recall; case-
2.0T EA888 | tensioner failure | warranty + $ | by-case goodwill
------------------+--------------------+----------------+------------------
Ford Focus/Fiesta | DPS6 PowerShift | 10yr/150k TCM | No recall; few
(2011-2016) | transmission fail | + buybacks | buybacks/payouts
------------------+--------------------+----------------+------------------
Hyundai / Kia | Theta II / Nu GDI | 3M+ recalled + | No mass recall
(2011-2019) | bearing/fire risk | lifetime wrty | (diff specs/plants)
------------------+--------------------+----------------+------------------
Tesla 3/Y/S/X | Autopilot nag & | 2M+ NHTSA OTA | No EU recall
(2021-2024) | UI font compliance | recalls | (UNECE rules)
sam1714 18 hours ago [-]
I'd also add the big driver for repairability is the US EPA which under Clean Air Act powers requires any emissions-related repair to be able to be done by third parties. They have mandated SAE J2534 for module updates and publicly available service manuals, for example.
Pure battery EVs are considered to have no emissions-related components, which is why Tesla gets away with that they do.
misterderpie 17 hours ago [-]
Thank you to both of you for letting me know.
SV_BubbleTime 14 hours ago [-]
J2534 is just a hardware interface.
You want to use it? Go ahead and buy a 1,3,7,30 day pass to the mfg’s service. Then, good luck with your J2534 because they only approve a couple, and better find a windows machine because their .NET program only runs on Windows 10.
thisismyanon2 13 hours ago [-]
J2534 codifies Win32.
dawnerd 1 days ago [-]
Which also has the real nifty benefit of a RRS via subscriptions!
knowaveragejoe 20 hours ago [-]
I'm hoping the slate truck will be different, and that whatever problems emerge out of giving consumers real control over the hardware they purchased doesn't hamstring the company somehow.
CamperBob2 24 hours ago [-]
They are now required to do just that, if they want to sell to the EU market. When you hear the term "cybersecurity," the real subject is the protection of revenue streams from attack by end users.
DANmode 21 hours ago [-]
Volkswagen shipped the same encryption key to the keys of something absurd like 100M vehicles.
Don’t hold your breath.
pontifier 19 hours ago [-]
This is why AI hype isn't overrated. When you are able to have an idea, and then use AI to move meaningfully in the direction of that goal so quickly we are going to see the entire landscape change. We really are nearing the singularity.
flatline 19 hours ago [-]
Not sure if I believe in the singularity. I think technology brings about singularities from time to time. Each one puts us on a new plateau which seemed unfathomable from the previous plateau but now looks commonplace. I realize the concept of singularity is that we accelerate unheeded to a point of infinite progress, but do you really think this current generation of LLMs (and this round of VC funding) is going to give me a wetware lab at home where I can give myself custom gene therapies to reverse disease and aging? The technology is cool but what I see at this point is most of its capability is already on display and people are rushing to carve out an ownership stake to secure profit. We already see copyright holders putting limits on LLM providers, safety measures, proprietary models solving the Big Problems. Wouldn’t surprise me if this kind of reverse engineering will be restricted to security researchers and black market ablated models in two years, unfortunately.
ftchd 18 hours ago [-]
There will be plenty open source models that would do reverse engineering, no black market or ablation needed.
The smarter the models get, the bigger the loss for the big 2 when an open source model gets released that doesn't talk down to you. The value of open source increases and closed source falls exponentially.
flatline 18 hours ago [-]
This is certainly what I am hoping for as well. I’m really curious to see where we are at in a year. We have reached a plateau where the labs clearly have capabilities out of reach of any public model. As far as I understand this is purely constrained by compute resources.
jrflo 18 hours ago [-]
The singularity has nothing to do with diffusion of the tech through society, the models can be superhuman in every capacity and only available inside Anthropic
dsign 2 hours ago [-]
> is going to give me a wetware lab at home where I can give myself custom gene therapies to reverse disease and aging?
For that specific end, you still money to buy the equipment, and a cabin with a basement, preferably in the woods on an island in the middle of the Pacific. Don't pick a tiny atoll that is going to be submerged in a few years. But other than that, since you don't need an army of scientists any longer, you can do with a smaller island and significantly less payroll expenses. What's not to like about the development?
ImaCake 14 hours ago [-]
Everyone mistakes sigmoids for exponentials until they pass the inflection point (see COVID for the classic recent example). I write this comment from a suburban house that exists only because we went through a "transportation singularity" which ended with fuel efficient cars and highways. No one talks about a transport time exponential where we will get to work in 1 second in 10 years time!
To protect "corporate interests" :( That article is about gaming, but this will only be the start of it of course.
saulr 18 hours ago [-]
I’ve seen this reported (and screenshots of it in action) but then earlier today Opus 5.5 used Ghidra to reverse engineered a bunch of logic from Counter-Strike 2. Is this being A/B tested?
AshleyGrant 17 hours ago [-]
I had it use Ghidra to review the firmware for a device I own just today. It started reviewing it and found some of the information I needed, but once I asked it something along the lines of "What happens when I press the B4 button if the device is in lockout mode?" (the thing I want to do is create a device that can send the same command as pressing that button but put my own button in a different location), Claude shut down completely.
So, I switched to a Chinese model to do the work. Anthropic and the US AI companies are trying to fight a tsunami with a teacup.
11 hours ago [-]
16 hours ago [-]
knowaveragejoe 20 hours ago [-]
If they won't, others will.
hmokiguess 23 hours ago [-]
The really interesting part to me is how he ends up by saying "I recommend you feed this repo into the agent and just have it help you set it up, it's super easy ..."
The average person is likely to feel overwhelmed by this, which I think is what is likely going to shape the new economy gig of "AI engineers". Those that aren't overwhelmed by it and that may not only enjoy doing this but also potentially figure out a way to scale operations that allow doing these kinds of DYI reverse engineer hack jobs.
sschueller 23 hours ago [-]
Aka the people who where manually engineering before and know how to use this new tool.
hmokiguess 23 hours ago [-]
It reminds me of the "handyman" economy in someways, it's gonna move towards "who wants to do it" and "who doesn't want to do it"
At least until the robots take over
cortesoft 20 hours ago [-]
For a while, but eventually that person will have some experience with an AI agent and realize how little they have to know to get it working. It only takes a few experiences before people get used to it.
aiiotnoodle 1 days ago [-]
Would love to see this being used to add keys to cars. A lot of cars now have tablets and screens that could be used to do this, but for some reason (money) I have to take my VW to the dealer...
I think the whole market for aftermarket keys could be completely eradicated with good software UX (and low pricing).
Doohickey-d 23 hours ago [-]
The whole aftermarket key ecosystem, even vehicles going back 20 years, is messy, horrible thing.
It's dominated by Chinese manufacturers of specialty tools (Autel, Xhorse, ...), which inevitably mandate bundled accounts, subscriptions, mandatory cloud features, etc. Even the older cars, you have either cheap Chinese tools, where you never know wheather it'll really work, or the manufacturer's tool which needs a dealer code, which of course the dealer won't have after 20 years.
I wrote about my 20 year old car-key-adding adventure here: https://jonas.bike/low-cost-diy-key-/-transponder-for-my-200... - it requires desoldering an eeprom, two pieces of hardware from aliexpress which come with pirated copies of software, and lots of hope that it'll work...
An open source community curated collection of scripts, which works with a cheap OBD dongle (which is all the chinese devices are: OBD hardware + per-car software), would be amazing.
aiiotnoodle 17 hours ago [-]
Great read and similar to the same route I went down. I also got the xhorse for a corsa, was able to program the transponder but wasn't able to get the lock/unlock central locking buttons working using xhorse. I ended up just living with it and sold the car with it as a "spare" key.
The corsa also needed a "security code" that wasn't with the log book like it was supposed to be and would have required a trip to the dealer to get butI read about ripping the eeprom off the back of the dashboard to get the code but decided to cut my losses.
londons_explore 22 hours ago [-]
Frustratingly those cheap OBD dongles (ELM327 clones) seem to be mostly designed for diagnostics. They effectively run a regex over all the canbus data and tell the computer what matches.
However they can't match and forward everything (not enough RAM) and miss loads of packets, which makes them useless for things like reflashing firmware or interfacing with OEM tooling.
holoduke 19 hours ago [-]
So what odb should I use to start reverse engineer my TCU from a old 2006 ferrari? These tcus are hard to come by. I wish I could replicate it's behavior. Not sure if that is possible. I think i can do it
londons_explore 18 hours ago [-]
In 2026 I would grab anything supporting SAE J2534 and then tell Claude to figure out all other details.
bri3d 22 hours ago [-]
Keys in European and newer American cars reliably have Some Kind of real-ish crypto scheme applied, as this is fairly regulated in Europe. It’s an anti theft measure driven by the real stolen auto trafficking issue faced especially in Europe. Most of the aftermarket tools vendors use some kind of exploit (code exec, arbitrary read, update signature bypass, etc.) to extract the crypto material from a control module participating in the immobilizer - it’s neither a weird conspiracy nor as simple as it seems to bypass.
terribleperson 23 hours ago [-]
I had this same thought the other day. For a lot of cars you can buy aftermarket programmers, but the Dorman ones are limited to 3 uses. 3 uses for like eighty dollars is terrible, and surely it's not a technical limitation.
catlifeonmars 22 hours ago [-]
Heh, maybe instead of circuit boards they are using wheat crackers.
terribleperson 22 hours ago [-]
Hah, I hadn't considered a hardware approach to limited use.
I assume there's just a counter in there. There could, theoretically, be cryptographic reasons it can only do 3 programming sessions, but I doubt it.
ronnier 21 hours ago [-]
You can do that with teslas. Buy two keys (they at those credit cards sized keys) and you pair them on your own. Two for $40. Do it right at home. You can pair multiple cars to a single key. Or make guest keys
XorNot 21 hours ago [-]
This works for most modern Toyotas - there's a sequence of door openings which put the car in pairing mode and will link a new key.
ronnier 20 hours ago [-]
But you have to buy an expensive key. I have two Teslas and a Toyota and my Toyota is so annoying that I even have to have a key. My Tesla just uses my phone (the credit card keys are totally optional)
trollbridge 1 days ago [-]
Good luck with Claude not just immediately giving you a refusal.
prodigycorp 1 days ago [-]
Codex would
ryandrake 22 hours ago [-]
It's crazy. Reverse engineering a protocol/interface is something LLMs are surprisingly good at, and these show a promise of unlocking sooo much proprietary crapware that's hidden behind obfuscation, accounts, subscriptions, lazy vendors who only write crappy windows programs, and so on.
But of course, the major US labs are going to get pressure from these vendors whose business models depend on obfuscation, and you know they're going to bend the knee and nerf the LLMs.
qingcharles 1 days ago [-]
Not these days. It won't even look at my projects for security flaws any longer. Has a little pop-up that says "This output cannot be viewed." or some such.
Have to make the Chinese one do it.
KlutzySofa 24 hours ago [-]
Its backwards in so many ways.
You fall back on having to use Chinese models. In theory the Chinese models could also be trained to sneak in new security flaws while Claude and Codex would avoid looking for them.
theplumber 24 hours ago [-]
Claude and codex could also be trained to sneak security flaws. They could even target specific people. Not to mention that all their data goes through NSA headquarters so you don’t even have the option to run it locally in any shape or form.
nullc 15 minutes ago [-]
Not just trained, but since you can only use them via an API they can have an external process decide when to add the bugs-- and use data that isn't available to an LLM (your user credentials, IP, information from parallel sessions or ad networks, potentially how long since you moved your mouse, depending on harness in use).
The API also can't be replayed for inspection-- if you do catch suspect behavior in an open weight model you can reproduce it.
Plus APIs can exflitrate your data through the API just by inspecting it... a maliciously trained open weight model needs to successfully make connections to do so.
KlutzySofa 23 hours ago [-]
The US companies are doing this to not get export restricted by the US government which in turn can end up with companies using Chinese models to get around those restrictions.
Its not really a matter of who's good and bad (neither the US or China has the best track record), but creating a scenario where the other party can use your own policies against you.
mrinf1nity 22 hours ago [-]
My lil' Claude tells me that my security related request is flagged and will be sent for human review. And to use a lesser model.
I told it 'Let's find the limitations of this Linux system' then it made a benchmarking set of tools for disk i/o, RAM, etc
When using Gemini on my phone it also told me it'll be 'human reviewed.' when asking it about car prices.
Also when saying 'let's use best security practices'
o_O
drnick1 22 hours ago [-]
Just tell it to continue. For now, it still mostly works. On the other hand, Claude flat out refuses to look at game and anti-cheat binaries, etc.
qingcharles 21 hours ago [-]
It sorta, sometimes does. It's not a reliable fix, though, sadly.
Even my old reliable Muse Spark has gotten a little testy lately. I'm back to GLM 5.3.
catlifeonmars 22 hours ago [-]
Get the Chinese one to jailbreak the proprietary one. Best of both worlds.
nullsanity 23 hours ago [-]
[dead]
zeryx 22 hours ago [-]
GLM or Qwen will do it
holoduke 19 hours ago [-]
It's pretty easy to bypass this check by telling the system you are working at a car Company and repair hardware with blabla. Many other tricks to get though those weak filters.
rasz 8 hours ago [-]
Especially keys to your car, by someone who plugged a dongle into headlight loom he got access to thru front wheel well.
rootsudo 14 hours ago [-]
The video is pretty inconsistent, the chrysler/french ECU has been disassembled before for a while, there are many aftermarket ECU's avaiable for the XJ cherokee, the carb/diagnostic use is normal for troubleshooting, the animation is painful to watch and makes presentation of the whole video....horrible, the narration clearly AI and horrible too.
people narrating and showing a video of what claude/an llm does just hurts my brain on watchability. what am I watching for?
someone else said it but yeah, meat proxy, zero effort.
d1l 23 hours ago [-]
Fascinating, but if he had low pressure on the rail that would point to a clog or possibly a pump issue, right? Wasn’t really clear whether the lean condition was due to the map or a mechanical issue.
I have a dirtbike and experienced some stuttering and stumbling under load after getting hot. Used Claude to help troubleshoot, as the plug was white and it was real lean. Replaced the pump (a 300$ module) and saved a ton of money and annoyance having a shop deal with it. 2000mi later zero issues. Absolutely game-changing. I use a dumb aftermarket fuel controller to add fuel under certain conditions, but would love to try and tune it myself some day.
iririririr 22 hours ago [-]
most likely, his new non-redundant non-operating-temperature-rated ECU is just turning on the fuel injectors, and completely ignoring all feedback the original ecu used. so until he hears the knocks, he will think he "fixed" it.
yolo engineering for ya.
ErroneousBosh 22 hours ago [-]
If you just turn the fuel injectors on all the time, they will overheat and fail very quickly and the engine won't run.
I've got a "get you home box" for EFI systems with separate ignition, like Bosch LE-Jetronic that I built a few of as clones of an insanely expensive device some 20 years ago. It really just consisted of a couple of 555 timers and a bloody great MOSFET, with one 555 wired as a Schmitt trigger picking up ignition pulses and the other delivering a 1ms pulse to the injector coils every time the spark fired. This dribbled enough fuel into the engine roughly proportionately to engine RPM to let you drive it slowly to somewhere you could work on it safely, or even just onto a trailer.
And of course, Speeduino, Megasquirt, and a bunch of others are a thing.
mianos 19 hours ago [-]
I made an obd reader https://github.com/mianos/canmqtt for my bike that is based on the Wavehare canbus board. The handy thing about the board is it already has buck converters for the bike power.
Once it's in the bike (I put it permanently in the tail) It's in the wifi and mqtt so I can point an llm at it.
muststopmyths 1 days ago [-]
can someone explain the root cause of the smog test failure to me ? it sounds like the car was using a California smog test-specific fuel map at the speeds of the test. Wouldn't that make it pass the test, not fail it ?
very cool project anyway.
threetonesun 1 days ago [-]
Yeah doesn’t really make sense, there’s very likely a mechanical issue the LLM “fixed” with a richer fuel mix. But also it’s a 94 XJ the reason the CA map was there was probably because they got abhorrent gas mileage without it.
eightysixfour 16 hours ago [-]
His particular car had a tune to reduce emissions by running lean at certain RPMs. While that reduces some emissions, it increases NOx, which is what was failing. You can add some more fuel to get closer to lambda, which will reduce the NOx and increase the CO and HC.
bflesch 24 hours ago [-]
You can adjust the mixing ratio of fuel and air in a way so that the exhaust has reduced emissions. This will reduce performance of the car but allows the exhaust to pass the emission tests.
schwap 18 hours ago [-]
In this case it's the opposite. High NOx readings are almost always an indicator of reduced performance.
tehlike 21 hours ago [-]
I have used Claude to build https://obd.wtf to essentially create a commercial grade diagnostic tool for cars. Now working on new version.
But don't use it, obd read functions are fine but other functions for other diagnostic/actuator etc stuff is not.
cannonpalms 19 hours ago [-]
Would love to contribute a dataset of Ford codes I have come across over the years. It's more complete than anything else out there for free.
tehlike 18 hours ago [-]
Oh wow. Can you email me at tehlike @ gmail.
I am taking this slow as I have other priorities, but as models advanced the stuff I Was able to do with limited time have increased, so with new version i still have locally, it may be possible to do more interesting stuff.
tehlike 18 hours ago [-]
Just making it clear: The intention was to make it commercial grade tool, but i am not even 1% there, only doing OBD as proof of concept, and everything else as "future work"
someothherguyy 19 hours ago [-]
what makes it "commercial grade"?
tehlike 18 hours ago [-]
Most / all consumer diagnostic tools are read only, often pull DTC codes from the vehicle. There are a ton of other stuff you can do with UDS for example. Maintenance resets, actuator testing, coding, all sorts of other stuff.
An example I recently did with my car was disabling auto start-top. not possible with consumer tools. Another simple one is maintenance resets - some consumer tools can do this. ABS bleeding - often you can't do it with consumer grade. etc.
rpcope1 14 hours ago [-]
It's an interesting idea, but I would strongly suggest being careful about offering active tests and procedures. The ABS bleeding as an example is possible if you're careful and understand the limitations with certain ELM327 clones or better yet an OBDLink or good J2534 device, but even if the procedure is coded correctly, something like Bluetooth has the risk of being disconnected, or if it's not a real J2534 setup you can have serious problems with timing or just getting a processor wedged. I could see an ABS module getting stuck open and broken this way, or some of the tests you could do in Techstream or Tech2 can legitimately damage a power train if you don't understand what you're doing. Programming via anything short of a real J2534 is also a siren song, as it's technically feasible theoretically even with something like a CANable or OBDlink MX+, but again timing, buffer sizes, and generally needing to provide external power basically anytime you need to do something like program (and that's not obvious if someone hasn't told you or you haven't bricked an ECM), would make me really hesitant to hand that gun to just anyone.
In thinking about it, someone is also likely to ask how you figured out how to do an active procedure like an ABS bleed. Maybe you sniffed an Autel or a Modus, but I wouldn't be surprised if certain manufacturers got up your ass very quickly about the provenence of your code. If you're in China like Autel was when they first built their tools, I guess you can ignore manufacturers being upset about the lost revenue stream, but otherwise I might be concerned.
tehlike 13 hours ago [-]
Fair points. Yes active comm needs a lot more testing & is fairly risky, so there needs to be very good guardrails (exact model matching for a given command etc), together with solid communication. I think bluetooth is much more reliable today than it used to be, but I wouldn't use a cheap elm device with this for sure.
tehlike 13 hours ago [-]
On a related topic, is there a place i can follow you (like on twitter etc).
lixtra 11 hours ago [-]
> 6:35 During the exact loads where we do the smog test the California computer trims the fuel down.
So Volkswagen like manipulation to pass tests dates back much further and even inner usa?
Edit: yes, volkswagens manipulation is not used in traffic, while this car behaves the same in real traffic.
KolmogorovComp 1 days ago [-]
The tech is impressive but the animation in the video are too, anyone knows what was used to generate them?
qingcharles 24 hours ago [-]
Definitely one of the Claude video skill stacks that are going around. Dude should make a whole separate video about the animations, he'd get as many views again just for that.
prodigycorp 1 days ago [-]
claude can do that lol
boznz 22 hours ago [-]
Spent so many wasted days reverse-engineering older systems to get them tto work again or prevent eWaste. In a perfect world all CAD files should become open source when the product becomes EOL
threecheese 18 hours ago [-]
Sonnet 5.5 threw a guardrail for me trying to create a test fixture to generate data in a dev environment with no PII.
How tf are you getting it to RE?
18 hours ago [-]
meandave 21 hours ago [-]
I use anthropic models for some of the diagnostic reads on https://crewchief.cc, it's been a lot of fun hacking on obd2 things that used to take much longer to decode before LLMs
blobbers 21 hours ago [-]
I’ve seen Claude run a cooler on a chip to -20 without saying anything. There was humidity in the air, and the condensation fried the board
after a few days. Thanks Claude. Maybe trying to keep the junction temps super low wasn’t such a bright idea.
prodigycorp 1 days ago [-]
Claude reverse engineered his car, drew the animations, and wrote the script for video. Incredibly weird meat proxy moment.
dghlsakjg 1 days ago [-]
It also didn’t solve the root cause of the issue. ECUs don’t need to be overwritten to get stock performance and emissions. Something is wrong with hardware/mechanical that needs to be solved.
What he did was technically interesting, but probably not a great solution to the actual problem.
d1l 22 hours ago [-]
Yeah I had the same impression. There was plenty of data presented and code was dumped and reversed…but was the actual issue solved? What even was causing the issue? Presumably it passed smog sometime in the past couple years so what degraded? It was lean and there was some kind of issue with pressure. I can’t imagine a reflash or whatever is the fix.
It definitely feels like a classic case of the operator reinforcing Claude to keep pursuing a goal that’s counterproductive to a robust solution. Claude complied and they went hard on it but I can’t imagine the driver is any better off and more likely in a worse spot as the underlying issue wasn’t solved.
mrbombastic 18 hours ago [-]
”You attach an assumption to a piece of evidence, you start to bend the narrative to support it. Prejudice yourself.”. I find by default while claude is great at analyzing logs and diagnostics from a variety of places it is really prone to this bad detective pattern of choosing an explanation and sticking with it even when evidence starts to bend away from it. you really have to play the role of the skeptic to do do good root cause analysis. Which is very hard to do in unfamiliar domains.
wildzzz 18 hours ago [-]
It's probably a bad MAF sensor but this guy thinks he knows more than he does and uses the king of incorrect confidence to try to diagnose the issue.
dghlsakjg 14 hours ago [-]
To be fair to AI, I put his symptoms into Claude and it suggested all of the obvious things and steps to start with. None of them were to remap the fuel delivery system.
slashdave 23 hours ago [-]
Yup. These agents have feedback loops that are self-reinforcing. Without technical background, a user is easily mislead.
teejaypeeps 24 hours ago [-]
And did the thumbnail, text and graphics. 1mn views with zero effort. I'm here for it
urig 21 hours ago [-]
100 points for car hacking fun. -60 points for "not needing a mechanic or anything". A good mechanic would have fixed it up in a half hour. Guy took 3 days and god knows how many tokens.
bitdivision 20 hours ago [-]
I don't believe a normal 'good' mechanic would be able to re-map it in 30 mins (assuming the underlying issue wasn't something else).
rootsudo 8 hours ago [-]
You don’t need to remap, on these inline 6 keep engines it most likely is cracked vacuum line or maf. Of all the measurements this guy pulled, he didn’t pull vacuum.
thisismyanon2 13 hours ago [-]
There was an underlying issue and if the state and/or feds caught you remapping the ECU you'd get smacked down hard.
schwap 18 hours ago [-]
There almost certainly IS an underlying issue, though. Unless the ecu spontaneously recapped itself.
rpcope1 14 hours ago [-]
It wasn't clear watching that video that a "remap" was remotely appropriate. A persistent +22% fuel trim, especially at a vehicle that age, is something like unmetered air, a stuck injector or dying fuel pump. Claude will confidently assert all sorts of absurd shit, when a normal mechanic would apply some common sense here. If you've never tried to setup a MegaSquirt, you have no idea how easy it is to screw up a fuel map and cause something like detonation. If anything, maybe you pull an ECM from a non-CARB XJ, but really a lean engine this age is going to be something way more pedestrian like cracked vacuum lines, a stuck injector, or this being an XJ, shitty wiring somewhere. At best I wonder if Claude is going to raise a new generation of parts changers.
diebillionaires 21 hours ago [-]
Mechanics are overpriced.
LeoPanthera 19 hours ago [-]
All those sites that currently recommend you do "curl | bash" are going to change to "curl | claude" and that scares me a little.
IshKebab 1 days ago [-]
Manufacturers are going to start cryptographically locking everything down rather than relying on the time and skill needed to do reverse engineering.
Rebelgecko 24 hours ago [-]
LLMs have been pretty good at breaking bad crypto in cars. In my Hyundai, they were able to figure out the right incantation to get the car to run unsigned code off a USB stick. Although I guess we're in trouble once carmakers use LLMs to lock things down even more
intrasight 23 hours ago [-]
Most shop security is security through obscurity. For them to lock it down, they need to manage those secrets all the way back to a cloud identity provider. They can and probably will do that eventually.
I assume Tesla does that already.
bri3d 22 hours ago [-]
Old news - for 20+ years in Europe, 5-10 years in the US, and now in Korean and Japanese cars too, ECM updates have been cryptographically signed for warranty, emissions, and now EU cybersecurity regulatory compliance. It turns out there are still bugs in things with cryptography too though.
tehlike 21 hours ago [-]
They did. Cayenne 2024+ does this and ecu is essentially read only unless you start a diagnostic session online.
muglug 20 hours ago [-]
Really cool, but what's the weather-sealing like on all that circuitry? He mentioned stuff flying around when you turn hard, which doesn't sound great.
matchagaucho 18 hours ago [-]
I'm almost afraid to look up the Kelly blue book value on this Jeep.
This guy clearly put $5K+ of time and video production effort into this cool hack.
winrid 12 hours ago [-]
he's getting pretty good marketing for himself out of it
rasz 8 hours ago [-]
Look at me, Im a meat (proxy) popsicle kind of marketing.
winrid 3 hours ago [-]
Pretty sure half the engineers vibe coding right now are adding less value than fixing an old Jeep.
I'd rather work with jeep guy than plenty of people in my career :D
EcommerceFlow 23 hours ago [-]
Love seeing unique youtube projects on here for a change!
pino83 23 hours ago [-]
... even better would be free projects.
redmaple892 22 hours ago [-]
I enjoyed watching this a couple days ago -- super cool as someone also into slightly older vehicles.
gameclimb 21 hours ago [-]
Oh hey I saw this post on Reddit, super cool!
truekonrads 16 hours ago [-]
This is anthropological gold.
varispeed 20 hours ago [-]
Reminds me when I dropped a binary file from an eeprom and asked Claude why device would display "Error.". It figured out what the processor was, how it talked to LCD and then from there how it displayed messages and then listed conditions for which Error would appear. Amazing.
russellbeattie 19 hours ago [-]
"Everything is 'open source' if you read machine code."
That's been a tired old saw in computers since my career began in the 1990s. Who knew it would be so applicable in the age of AI?
rasz 8 hours ago [-]
"How I and Claude cheated California smog test, Im sure nothing bad will happen to me nor the station owner who witnessed this fuckery"
> Claude suggested that it would run the fuel pump on its own. Um and I said, "Hey, if you can do that, why not? I don't think you can do that." And it started priming the fuel rail and it kept priming it and kept priming it. The fuel pump was on for about 10 seconds and I started to freak out uh and requested it to stop because that was the first time that Claude had actually done something inside of the computer to have a change.
Although we've seen this play out with handing cloud keys to Claude and it just going off on its own, seeing it poke around and make physical, real-world changes in its clumsy, cavalier way is both awesome and terrifying.
I am not giving Claude access to physical objects until we can reliably contain it. It also seems to me that applying software permissions logic to interactions with the physical world will backfire badly.
The fuel pump runs at whatever PSI, and most of the fuel isn't used and gets dumped by the regulator back into the tank.
I dunno about 94 jeeps, but the fuel pump is probably not modulated, and the pressure regulator may well be vacuum modulated... Running the engine hard, get more fuel pressure. If the diaphram fails, get fuel into the vacuum lines (oops).
As long as the regulator is dumping fuel to the return line, you're good. If flow stops, your fuel pump stops getting liquid cooling and the bearings probably wear out. If the fuel runs into the vacuum lines and you flood the engine, that's not great, but probably not catastrophic, unless you ignore the fuel smells.
You are very clearly wildly unqualified to be "advising" anyone on anything about cars.
I think they should be held fully accountable for every cent of damage caused by models they ran and gave either the tools to hack, instructions to hack, or incentives to hack.
In every AGENTS.md/README.md for a component that touches hardware (and in the root level) we have instructions to never run hardware or touch running code without explicit permission.
It will happily build its own jail, and (at least assuming you're not using hobbled US commercial models) it will test if the jail can hold it too.
If you don't want the machine to move, unplug it or press the emergency stop.
For example “never, ever put milk in bowl, ever, under any circumstances” and then someone asked it for a bowl of Cheerios, and it took that as an instant approval to ignore that rule and not even let us know. And when we asked why it said “who would each Cheerios without milk?”
And for the love of humanity, don’t give it access to the hosts docker socket! It will figure that out faster than you can read it…
In general I've seen them split between 30% brilliance, and 70% pants of head outrageously stupid shit for electronics and automotive stuff (both diagnostics and repair/modification). I would not suggest even frontier models if someone didn't understand what they were doing.
As a Codex user, what do you mean? Does Claude actually try to report issues back to Anthropic when something goes wrong?!
I assume the corpus of text they were trained on matches that 30/70 split?
I assume there's some selection or fine-tuning that happens to improve models fed similarly unreliable training data for other domains... curious how this works.
For whatever it's worth, that a) Astra struggles less on this for some reason, and b) things tend to work better if you have a copy of the FSM and some other appropriate lit like Halderman and quality forum posts or alldata if it's appropriate and tell it to work from there. I've been working on a big search engine for this that feeds a known good corpus like an ass load of contemporary MotorAge mags, relevant technician texts, the FSM, other quality data sources like BITOG, known not-shit forums, along with OBD2 data via something like an OBDLink dongle or a J2534 setup. I still would not trust it, but it can be helpful to eliminate or brainstorm some possibilities.
Factory service manual, in case anyone else is wondering what the finite state machine in your car is for…
finnicking... pondering... clauding..
529 please try again later
This isn't meant as a defense of Claude or to refute the OP anecdote, but rather to point out that all models are non-deterministic even when safeguards are meant to exist.
Just because they don't do the stupid thing 90% of the time doesn't mean they won't do the stupid thing 10% of the time.
So when tasks are actually dangerous one should use models assuming they will do the stupid thing 100% of the time and find your own ways of avoiding the damage with that assumption in mind.
I notice with mine that it only matches my level, by default it's a dump slop generator, then if I start asking, "shouldn't you test this, shouldn't you refactor that" suddenly it 'remembers' all kinds of software engineering knowledge.
Otherwise I've been pretty happy with it. It's been absolute blast with Home Assistant and OpenWRT (tightening up network security and cutting off iot devices from WAN).
Even if AI goes away tomorrow, the tools will still exist.
Is this just "vibes"? Maybe, but I now have a full, verified spec, against which to build software for myself, also with a little bit of "vibes".
Is this a good or bad thing?
that is entirely dependent how accurate the spec is
> verified spec
I'm currently teasing apart numerous ancient Text to Speech synthesizers with Sol and the amount of times I ask a follow up question that forces it to correct 5 things it's already said/documented is ridiculous.
What degree of verification would make you say it's worth something?
But this is your unspecified expensive device. A verified spec wouldn't have you send a command or configure a table that bricks or damages it. A vibe verified spec very well might. How much is that worth to you? What are you willing to lose?
Hard to judge though, most people are just winging it though and shoving it out the door as if its impressive.
But yes, I think its probably bad because you probably didnt learn all that much and became more reliant on Anthropic. Unless you dont care about learning, then who cares.
Same with my bespoke IRC client, RSS feed hydrators and a dozen other small utilities I've vibe coded. All stuff that I've been planning for a long time, but haven't found out the time to dedicate to them.
Now I'm saving actual money every year because I could drop subscriptions applications.
Not every application needs a weekly release with new groundbreaking features. They're just one and done. Maybe an update if you hit a sharp edge while using it.
It's not a massively complicated thing, just a single server that says online 24/7, which talks using a custom protocol to different clients (Web, CLI, Linux native and macOS native).
I could've used pre-existing software to do most of it, but there's always jank and things I can't customise without trying to get changes into the mainline.
There are many IRC clients and this one is mine. I'm not expecting anyone else to use it. It won't have a community, a snazzy web page and it won't be taking in any issues or PRs from anyone =)
I have used pretty much all of them and they’re all just that little bit off from what I want. Quassel, Lounge, Irssi etc
Forking is a bunch of work I don’t want. As is trying to get specific changes merged to main.
I have been using my videcoded clone of Adobe Illustrator for a while now.
No intentions of releasing/selling it, as I'm sure everyone else can now do similar things too.
Just because you've chosen to turn off your brain and prompt your way to whatever, and your 10 MAU app works fine does not mean large systems, critical devices and critical infrastructure's code will no longer need reviewed. Sorry your dad's pacemakers firmware code, and the ancillary systems that enable its development will need its code reviewed. Along with aircraft, satellites, municiple traffic systems, countless medical devices ect ect. I could go on forever.
Not everything is non serious SaaS platforms with little users. Just because you have success writing slop that exists under no load, and nothing relies on it doesn't mean the world is now different.
You have SEVERE dunning kruger syndrome, and unfortunately a lot of other people have this same ailment and the worlds systems are decaying as everyone turns off their brain and offshores everything difficult to language models, because humans like you are lazy and will always try to get the free lunch that doesn't exist. People will die because of this attitude.
Also I know at least one satellite company who definitely uses AI-assisted programming (I interviewed for an AI-heavy position in there).
Also in the mobile gaming industry AI is not even optional nowadays, every single AAA household name game has "AI" in it at every level - some used AI when it was called just Machine Learning.
I suspect you're right BUT still maintain a bit of optimism. We miiiight end up in a future where most hardware comes with an OK-ish app/interface and some sort of LLM-Friendly technical documentation so if you don't like $aspect of the device, you're free to roll your own after indemnifying the manufacturer via something like an eFuse.
or maybe I really am naive :(
I think most people on here fail to realize the demand to customize and build things hasnt gone up. Just because llms exist now doesnt mean normies all of a sudden want to start getting into programming, they dont care.
The only number thats gone up is people who were slightly interested in programming/CS before but were too undisciplined to learn things themselves. Llms are great for those lazy people.
I still get the joy of translating ideas and orchestrating and architecting solutions to problems I have in real life, but just at a quicker pace.
If some of those tools and solutions can help other people, that's great. If not, that's okay too. I hope everyone sort of has that mindset who's building today.
Me!
haha I dont even know github I use it as google
Yes I have full time IT job xD
stopping it would require outright putting DRM on the GPU and any other hardware that can run it, good luck with that (and not having the LLM itself be used to break out of it)
(to stay within physical machinery, like the post.)
actually I have no idea what would llm do to make their machines better
people were able to revive, repair, optimize a lot of day to day devices, very nice view. that said cars and heavy appliances.. should be done in a big public warehouse safety lab
Claude is quick to refuse to help with reverse engineering and I suspect it's only going to get worse over time.
As far as I can tell no manufacturer disables the car yet when detecting modifications, or take drivers to court for doing it, yet.
That's true but we have the benefit of unlimited time, having physical control, the analog hole.
> As far as I can tell no manufacturer disables the car yet when detecting modifications, or take drivers to court for doing it, yet.
As far as I know cars are a special case in much of the world (US, EU, others?). Right to repair laws laws usually give you the right to use non-OEM parts.
Aren't DMCA anti-circumvention provisions the bigger concern in the US? I think the EU has an exemption to allow reverse engineering for the purposes of interoperability and repair.
We do, but everything is "software defined car" now and loopholes usually get closed with mandatory software updates done with the mandatory inspections. They put everything in one basket together with safety critical systems because that justifies any lockdown.
> Right to repair laws laws usually give you the right to use non-OEM parts.
Manufacturers already make that very expensive. Mechanics have to be certified, they need special and very expensive tools, everything is wrapped in cryptography and the manufacturer has to be in the middle of everything, for a price. [0] So go ahead, use your OEM part, you'll still pay thousands for the mechanic that has the necessary prerequisites and for the manufacturer's services.
I pray I'm wrong but we need strong regulation and enforcement, and the manufacturers can still maliciously comply.
[0] https://news.ycombinator.com/item?id=46155619
Nothing prevents you from unplugging the cell modem. The car will still work.
Why would they? They just need to fund lobbyist to make it illegal for any AI to reverse engineer anything copyrighted or with IP.
Sure, you can use open source AI to do it but not everyone will/can do it.
So, they don't need to ban everyone, they just need to make sure the final code isn't available freely on GitHub. That would minimize the damage, it's not a perfect solution but more than enough to protect the company.
Making something impossible to do is more thorough than making it illegal to do
Unless your company's moat is in the thing that was reversed engineer, I don't think most company cared enough to spend huge amount of resource to prevent it technically. The only thing that they need to prevent is the liability that comes after the product is used inappropriately or illegally.
And it you can defend something with AI, there's no reason why you can't attack it AI.
Lockstep processors, hardware security modules with immutable unreadable keys, and PKI keys that the mfg holds the private for that can only be accessed on their online services.
It’s locked down now and this started in 2018 or so. It’s just worse now.
To automotive, there is really still no right to repair. They get around it by selling “3 day passes” to their service, but with their hardware and at their prices.
They’re all moving the Tesla system of more centralized brain systems which will be even more protected.
Let's be clear that it's just incentives. If the EU economy was based on creating companies and new products there would not be as many incentives for consumer "rights" there.
Not for car owners/consumers in the EU. Auto makers and dealers get away with a lot of shady shit fucking over their customers in the EU all the time. For example we in the EU don't have the same trigger happiness of NHTSA mass recalls the Americans have, but matters of such auto issues/defects is left to individual member states or even at individual case by case to decide what action is needed, and responses are a lot weaker and more limp-dick than in the US, almost never leading to mass recalls, even for the same issues[1], mainly because the US is more litigious making car manufacturers to be on the edge and proactive to issues.
Like sibling said, there are incentives to regulate stuff that has no impact on EU businesses, shit like USB-C or replaceable batteries, but they won't touch the auto industry to mandate easier repairs and less lock-downs.
Pure battery EVs are considered to have no emissions-related components, which is why Tesla gets away with that they do.
You want to use it? Go ahead and buy a 1,3,7,30 day pass to the mfg’s service. Then, good luck with your J2534 because they only approve a couple, and better find a windows machine because their .NET program only runs on Windows 10.
Don’t hold your breath.
The smarter the models get, the bigger the loss for the big 2 when an open source model gets released that doesn't talk down to you. The value of open source increases and closed source falls exponentially.
For that specific end, you still money to buy the equipment, and a cabin with a basement, preferably in the woods on an island in the middle of the Pacific. Don't pick a tiny atoll that is going to be submerged in a few years. But other than that, since you don't need an army of scientists any longer, you can do with a smaller island and significantly less payroll expenses. What's not to like about the development?
To protect "corporate interests" :( That article is about gaming, but this will only be the start of it of course.
So, I switched to a Chinese model to do the work. Anthropic and the US AI companies are trying to fight a tsunami with a teacup.
The average person is likely to feel overwhelmed by this, which I think is what is likely going to shape the new economy gig of "AI engineers". Those that aren't overwhelmed by it and that may not only enjoy doing this but also potentially figure out a way to scale operations that allow doing these kinds of DYI reverse engineer hack jobs.
At least until the robots take over
I think the whole market for aftermarket keys could be completely eradicated with good software UX (and low pricing).
It's dominated by Chinese manufacturers of specialty tools (Autel, Xhorse, ...), which inevitably mandate bundled accounts, subscriptions, mandatory cloud features, etc. Even the older cars, you have either cheap Chinese tools, where you never know wheather it'll really work, or the manufacturer's tool which needs a dealer code, which of course the dealer won't have after 20 years.
I wrote about my 20 year old car-key-adding adventure here: https://jonas.bike/low-cost-diy-key-/-transponder-for-my-200... - it requires desoldering an eeprom, two pieces of hardware from aliexpress which come with pirated copies of software, and lots of hope that it'll work...
An open source community curated collection of scripts, which works with a cheap OBD dongle (which is all the chinese devices are: OBD hardware + per-car software), would be amazing.
The corsa also needed a "security code" that wasn't with the log book like it was supposed to be and would have required a trip to the dealer to get butI read about ripping the eeprom off the back of the dashboard to get the code but decided to cut my losses.
However they can't match and forward everything (not enough RAM) and miss loads of packets, which makes them useless for things like reflashing firmware or interfacing with OEM tooling.
But of course, the major US labs are going to get pressure from these vendors whose business models depend on obfuscation, and you know they're going to bend the knee and nerf the LLMs.
Have to make the Chinese one do it.
You fall back on having to use Chinese models. In theory the Chinese models could also be trained to sneak in new security flaws while Claude and Codex would avoid looking for them.
The API also can't be replayed for inspection-- if you do catch suspect behavior in an open weight model you can reproduce it.
Plus APIs can exflitrate your data through the API just by inspecting it... a maliciously trained open weight model needs to successfully make connections to do so.
Its not really a matter of who's good and bad (neither the US or China has the best track record), but creating a scenario where the other party can use your own policies against you.
I told it 'Let's find the limitations of this Linux system' then it made a benchmarking set of tools for disk i/o, RAM, etc
When using Gemini on my phone it also told me it'll be 'human reviewed.' when asking it about car prices.
Also when saying 'let's use best security practices'
o_O
Even my old reliable Muse Spark has gotten a little testy lately. I'm back to GLM 5.3.
people narrating and showing a video of what claude/an llm does just hurts my brain on watchability. what am I watching for?
someone else said it but yeah, meat proxy, zero effort.
I have a dirtbike and experienced some stuttering and stumbling under load after getting hot. Used Claude to help troubleshoot, as the plug was white and it was real lean. Replaced the pump (a 300$ module) and saved a ton of money and annoyance having a shop deal with it. 2000mi later zero issues. Absolutely game-changing. I use a dumb aftermarket fuel controller to add fuel under certain conditions, but would love to try and tune it myself some day.
yolo engineering for ya.
I've got a "get you home box" for EFI systems with separate ignition, like Bosch LE-Jetronic that I built a few of as clones of an insanely expensive device some 20 years ago. It really just consisted of a couple of 555 timers and a bloody great MOSFET, with one 555 wired as a Schmitt trigger picking up ignition pulses and the other delivering a 1ms pulse to the injector coils every time the spark fired. This dribbled enough fuel into the engine roughly proportionately to engine RPM to let you drive it slowly to somewhere you could work on it safely, or even just onto a trailer.
And of course, Speeduino, Megasquirt, and a bunch of others are a thing.
very cool project anyway.
But don't use it, obd read functions are fine but other functions for other diagnostic/actuator etc stuff is not.
I am taking this slow as I have other priorities, but as models advanced the stuff I Was able to do with limited time have increased, so with new version i still have locally, it may be possible to do more interesting stuff.
An example I recently did with my car was disabling auto start-top. not possible with consumer tools. Another simple one is maintenance resets - some consumer tools can do this. ABS bleeding - often you can't do it with consumer grade. etc.
In thinking about it, someone is also likely to ask how you figured out how to do an active procedure like an ABS bleed. Maybe you sniffed an Autel or a Modus, but I wouldn't be surprised if certain manufacturers got up your ass very quickly about the provenence of your code. If you're in China like Autel was when they first built their tools, I guess you can ignore manufacturers being upset about the lost revenue stream, but otherwise I might be concerned.
So Volkswagen like manipulation to pass tests dates back much further and even inner usa?
Edit: yes, volkswagens manipulation is not used in traffic, while this car behaves the same in real traffic.
How tf are you getting it to RE?
What he did was technically interesting, but probably not a great solution to the actual problem.
It definitely feels like a classic case of the operator reinforcing Claude to keep pursuing a goal that’s counterproductive to a robust solution. Claude complied and they went hard on it but I can’t imagine the driver is any better off and more likely in a worse spot as the underlying issue wasn’t solved.
I assume Tesla does that already.
This guy clearly put $5K+ of time and video production effort into this cool hack.
I'd rather work with jeep guy than plenty of people in my career :D
That's been a tired old saw in computers since my career began in the 1990s. Who knew it would be so applicable in the age of AI?